[HDFS-10761] libhdfs++: Fix broken logic in HA retry policy

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Component/s: hdfs-client
• Assignee: James Clampffer
• Reporter: James Clampffer
• Created: 15/Aug/16 16:27
• Updated: 22/Mar/18 21:57
• Resolved: 22/Aug/16 21:36

Description

Two issues in the HA policy:

1) There's logic to guard against HDFS-8161, but it wasn't checking how many failovers had already happened. It'd get stuck alternating between nodes forever.

2) Switched a ">" for a "<" so when max failovers was reached it'd keep trying anyway.